User Behavior Extraction Device for Advertisement Targeting

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing advertisement distribution technologies struggle to accurately identify and target users with similar web browsing habits, leading to ineffective advertisement delivery as they fail to quantify the degree of similarity between users based on their browsing history, resulting in advertisements being distributed to uninterested users.

Innovation Solution

An extraction device and method that acquires user behavior histories and extracts target users expected to perform specific behaviors by generating a user behavior prediction model based on designated behavior histories and explanation behaviors, incorporating advertiser knowledge to determine similarity and improve advertisement effectiveness.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If advertisement distribution is performed based on user information such as preferences, gender, age, address, or occupation, then targeted distribution can be achieved, but it is difficult to accurately identify users with similar web browsing habits and quantify the degree of similarity between users

Engineering Contradiction:

Improveaccuracy of user similarity identificationVSAvoidinability to quantify degree of similarity

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The patent transforms the abstract concept of user similarity into quantifiable parameters by extracting specific behavior history items (e.g., number of pages viewed, time spent on sites, click patterns) and converting them into numerical data. This allows the degree of similarity to be measured and compared objectively, resolving the problem of unable to quantify user similarity.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces subjective judgment of user similarity with an automated information processing system that uses algorithms to calculate similarity degrees based on behavior history data. This substitution of manual/subjective methods with automated computational methods enables precise measurement and quantification of user similarity.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Quantity of substance

If advertisement is distributed to identified groups of users with similar web browsing habits, then the audience size is increased, but advertising effectiveness is not always increased because users may not be accurately identified

Engineering Contradiction:

Improvenumber of advertisement recipientsVSAvoidaccuracy of target user identification

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The patent performs preliminary analysis and extraction of behavior history items before advertisement distribution. By pre-processing user data to identify key behavior patterns and calculating similarity degrees in advance, the system prepares accurate target user lists before the actual advertisement distribution, ensuring both large reach and high accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent incorporates feedback mechanisms where the results of advertisement distribution are used to refine and improve the similarity calculation algorithms. By continuously learning from actual user responses to advertisements, the system improves its ability to accurately identify target users while maintaining large audience reach.

Inventive Principle:

Principle #23Feedback

3Device complexity

If only web browsing history is used to identify similar users, then the identification process is simple, but it is hard to ensure that similar users are accurately identified

Engineering Contradiction:

Improvesimplicity of user identification processVSAvoidaccuracy of user similarity measurement

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments the user identification process into distinct stages: data collection, behavior history extraction, similarity calculation, and target user selection. This segmentation allows each stage to be optimized independently, maintaining overall process simplicity while improving measurement precision through specialized processing at each step.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS11455660B2Extraction device, extraction method, and non-transitory computer readable storage medium

Publication Date: 2022.09.27 YAHOO JAPAN CORP

AI summary

The extraction device according to the present application includes an acquisition unit and an extraction unit. The acquisition unit acquires behavior histories of users being candidates to whom content is to be distributed. The extraction unit extracts target users expected to perform specified behavior, based on a behavior history designated by a content provider, of the behavior histories acquired by the acquisition unit. For example, the extraction unit uses a model for determining the degree of similarity between a user performing the specified behavior and a target user expected to perform specified behavior based on the behavior history designated by the content provider in order to extract the target users.